In this project we will develop an innovative intervention for preventing drug abuse and other unhealthy behaviors. At-risk students share two readily identifiable and school-relevant characteristics - they do not have a strong sense of attachment to school and they do not have positive friends. These characteristics are also strongly related to other risk and protective factors. Previous research demonstrates that congregating high-risk youth only increases their risk and deviance. We propose a strategy that will distribute at-risk students among low-risk groups in a way that maximizes the potential for them to develop friendships with prosocial peers. We will develop a social network intervention strategy that inserts at-risk students into small groups of students increasing their chances of positive social acceptance. In this R34 project we will: 1. Refine a social network survey that will assess students' network of within-grade friends, bonding to school, interests, and values. 2. Develop computer applications that will allow us to test a method for identifying at-risk students. 3. Develop computer applications that will allow us to test alternative methods for creating low-risk groups. (At least four are to be considered - Clustering Nearest Friends, Grouping using the Newman-Girvan algorithm, Weak-Link Clustering, and High-Centrality Seeding.) 4. Develop computer applications that will allow us to test alternative methods for inserting at-risk students into low-risk groups. (At least two are to be considered - Matching Group Scores and Matching Individual Scores.) 5. Collect pretest data from students. 6. Test each alternative for identifying at-risk students, creating low-risk groups, and inserting at- risk students into low-risk groups, comparing the effects of each in order to identify optimal methods for constructing social network intervention groups. 7. Conduct a pilot study of the intervention. We will provide one group of schools with intervention group lists and encourage them to convene groups as often as possible to provide opportunities for social interaction. A second group of schools will serve as delayed-treatment controls. 8. Posttest students and assess the relative effectiveness of the intervention for creating positive friendships among at-risk students and improving their bonding to school. Students in eight middle schools will complete surveys in the spring of two consecutive school years. Data collected during the first year will be used to test at-risk student identification, low-risk group formation, and at-risk student insertion algorithms. Four middle schools will adopt the team assignment intervention during the following school year. Students in all eight middle schools will again be surveyed at the end of the second year. We hypothesize that at-risk students in intervention schools will have more friends, will report increased bonding to school, and will have better attendance, less truancy, less school misbehavior, and improved academic performance compared to at-risk students in control schools. If successful, this intervention will be tested for efficacy of preventing drug use in subsequent randomized control research trials.